We propose a theoretical model that successfully describes numerically such major features of the electrical conductivity of the surface of ice as its magnitude and frequency dependence. Comparison of the model with experimental results on sigma(s) confirms that the magnitudes of the ice surface charge density and ice surface potential are close to those observed in Independent measurements. The model predicts the thickness of an electric double layer which is comparable with the experimentally observed thickness of the liquidlike layer on ice (similar to 5 nm at -10 degrees C). This coincidence raises a question about the role of electrostatic pressure and possible superionic transition in the specific structure and properties of the surface of ice.